Increasing interest in computer-based surgical simulators as time- and cost-efficient training tools has introduced a new problem: objective evaluation of surgical performance based on scoring metrics provided by surgical simulators. This project employed fuzzy set theory to design a classifier for performance of a subject training on a surgical simulator, using three categories: novice, intermediate, and expert.

The MIST-VR simulator was used in a user study of 26 subjects with three different surgical skill levels: 8 experienced laparoscopic surgeons (experts), 8 surgical assistants (intermediates), and 10 nurses (novices). Subjects were required to perform four trials of a suturing task and a knot-tying task on the simulator. The performance data were then used to train and test two fuzzy classifiers for each task. The fuzzy classifier was able to classify the users of the system. The models presented a highly nonlinear relationship between the inputs (performance metrics) and output (fuzzy score) of the system, which may not be effectively captured with classical classification approaches. Fuzzy classifiers, however, can offer effective tools to handle the complexity and fuzziness of objective evaluation of surgical performances.

This content is only available as a PDF.